Machine learning for Gravity Spy

Glitch classification and dataset

S. Bahaadini*, V. Noroozi, N. Rohani, S. Coughlin, M. Zevin, J. R. Smith, Vicky Kalogera, Aggelos K Katsaggelos

*Corresponding author for this work

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The detection of gravitational waves with ground-based laser-interferometric detectors requires sensitivity to changes in distance much smaller than the diameter of atomic nuclei. Though sophisticated machinery and techniques have been developed over the past few decades to isolate such instruments from non-astrophysical noise, the detectors are still susceptible to instrumental and environmental noise transients known as “glitches,” which hinder searches for transient gravitational waves. The Gravity Spy project is an effort to comprehensively classify the glitches that afflict gravitational wave detectors into morphological families by combining the strengths of machine learning algorithms and citizen scientists. This paper presents the initial Gravity Spy dataset used for citizen scientist and machine learning classification – a static, accessible, documented dataset for testing machine learning supervised classification. Previous versions of this dataset used in [8, 53] did not include all current classes and also for some of the classes, some samples were pruned and added. This set consists of time–frequency images of LIGO glitches and their associated metadata. These glitches are organized by time–frequency morphology into 22 classes for which descriptions and representative images are presented. Results from the application of state-of-the-art supervised classification methods to this dataset are presented in order to provide baselines for future glitch classification work. Standard splitting for training, validation, and testing sets are also presented to facilitate the comparison between different machine learning methods. The baseline methods are selected from both traditional and more recent deep learning approaches. An ensemble framework is developed that demonstrates that combining various classifiers can yield a more accurate model for classification. The ensemble classifier, trained with the standard training set, achieves 98.21% accuracy on the standard test set.

Original languageEnglish (US)
Pages (from-to)172-186
Number of pages15
JournalInformation Sciences
Volume444
DOIs
StatePublished - May 1 2018

Fingerprint

Learning systems
Gravity
Machine Learning
Gravitation
Gravitational Waves
Gravity waves
Supervised Classification
Detector
Baseline
Detectors
Ensemble Classifier
Classifiers
Testing
Test Set
Metadata
Nucleus
Learning Algorithm
Ensemble
Classify
Classifier

Keywords

  • Classification
  • Dataset
  • Deep learning
  • Gravity Spy
  • Machine learning
  • aLIGO

ASJC Scopus subject areas

  • Software
  • Control and Systems Engineering
  • Theoretical Computer Science
  • Computer Science Applications
  • Information Systems and Management
  • Artificial Intelligence

Cite this

Bahaadini, S., Noroozi, V., Rohani, N., Coughlin, S., Zevin, M., Smith, J. R., ... Katsaggelos, A. K. (2018). Machine learning for Gravity Spy: Glitch classification and dataset. Information Sciences, 444, 172-186. https://doi.org/10.1016/j.ins.2018.02.068
Bahaadini, S. ; Noroozi, V. ; Rohani, N. ; Coughlin, S. ; Zevin, M. ; Smith, J. R. ; Kalogera, Vicky ; Katsaggelos, Aggelos K. / Machine learning for Gravity Spy : Glitch classification and dataset. In: Information Sciences. 2018 ; Vol. 444. pp. 172-186.
@article{31ffca1e82e94ed797d33e02a8a36dff,
title = "Machine learning for Gravity Spy: Glitch classification and dataset",
abstract = "The detection of gravitational waves with ground-based laser-interferometric detectors requires sensitivity to changes in distance much smaller than the diameter of atomic nuclei. Though sophisticated machinery and techniques have been developed over the past few decades to isolate such instruments from non-astrophysical noise, the detectors are still susceptible to instrumental and environmental noise transients known as “glitches,” which hinder searches for transient gravitational waves. The Gravity Spy project is an effort to comprehensively classify the glitches that afflict gravitational wave detectors into morphological families by combining the strengths of machine learning algorithms and citizen scientists. This paper presents the initial Gravity Spy dataset used for citizen scientist and machine learning classification – a static, accessible, documented dataset for testing machine learning supervised classification. Previous versions of this dataset used in [8, 53] did not include all current classes and also for some of the classes, some samples were pruned and added. This set consists of time–frequency images of LIGO glitches and their associated metadata. These glitches are organized by time–frequency morphology into 22 classes for which descriptions and representative images are presented. Results from the application of state-of-the-art supervised classification methods to this dataset are presented in order to provide baselines for future glitch classification work. Standard splitting for training, validation, and testing sets are also presented to facilitate the comparison between different machine learning methods. The baseline methods are selected from both traditional and more recent deep learning approaches. An ensemble framework is developed that demonstrates that combining various classifiers can yield a more accurate model for classification. The ensemble classifier, trained with the standard training set, achieves 98.21{\%} accuracy on the standard test set.",
keywords = "Classification, Dataset, Deep learning, Gravity Spy, Machine learning, aLIGO",
author = "S. Bahaadini and V. Noroozi and N. Rohani and S. Coughlin and M. Zevin and Smith, {J. R.} and Vicky Kalogera and Katsaggelos, {Aggelos K}",
year = "2018",
month = "5",
day = "1",
doi = "10.1016/j.ins.2018.02.068",
language = "English (US)",
volume = "444",
pages = "172--186",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",

}

Machine learning for Gravity Spy : Glitch classification and dataset. / Bahaadini, S.; Noroozi, V.; Rohani, N.; Coughlin, S.; Zevin, M.; Smith, J. R.; Kalogera, Vicky; Katsaggelos, Aggelos K.

In: Information Sciences, Vol. 444, 01.05.2018, p. 172-186.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Machine learning for Gravity Spy

T2 - Glitch classification and dataset

AU - Bahaadini, S.

AU - Noroozi, V.

AU - Rohani, N.

AU - Coughlin, S.

AU - Zevin, M.

AU - Smith, J. R.

AU - Kalogera, Vicky

AU - Katsaggelos, Aggelos K

PY - 2018/5/1

Y1 - 2018/5/1

N2 - The detection of gravitational waves with ground-based laser-interferometric detectors requires sensitivity to changes in distance much smaller than the diameter of atomic nuclei. Though sophisticated machinery and techniques have been developed over the past few decades to isolate such instruments from non-astrophysical noise, the detectors are still susceptible to instrumental and environmental noise transients known as “glitches,” which hinder searches for transient gravitational waves. The Gravity Spy project is an effort to comprehensively classify the glitches that afflict gravitational wave detectors into morphological families by combining the strengths of machine learning algorithms and citizen scientists. This paper presents the initial Gravity Spy dataset used for citizen scientist and machine learning classification – a static, accessible, documented dataset for testing machine learning supervised classification. Previous versions of this dataset used in [8, 53] did not include all current classes and also for some of the classes, some samples were pruned and added. This set consists of time–frequency images of LIGO glitches and their associated metadata. These glitches are organized by time–frequency morphology into 22 classes for which descriptions and representative images are presented. Results from the application of state-of-the-art supervised classification methods to this dataset are presented in order to provide baselines for future glitch classification work. Standard splitting for training, validation, and testing sets are also presented to facilitate the comparison between different machine learning methods. The baseline methods are selected from both traditional and more recent deep learning approaches. An ensemble framework is developed that demonstrates that combining various classifiers can yield a more accurate model for classification. The ensemble classifier, trained with the standard training set, achieves 98.21% accuracy on the standard test set.

AB - The detection of gravitational waves with ground-based laser-interferometric detectors requires sensitivity to changes in distance much smaller than the diameter of atomic nuclei. Though sophisticated machinery and techniques have been developed over the past few decades to isolate such instruments from non-astrophysical noise, the detectors are still susceptible to instrumental and environmental noise transients known as “glitches,” which hinder searches for transient gravitational waves. The Gravity Spy project is an effort to comprehensively classify the glitches that afflict gravitational wave detectors into morphological families by combining the strengths of machine learning algorithms and citizen scientists. This paper presents the initial Gravity Spy dataset used for citizen scientist and machine learning classification – a static, accessible, documented dataset for testing machine learning supervised classification. Previous versions of this dataset used in [8, 53] did not include all current classes and also for some of the classes, some samples were pruned and added. This set consists of time–frequency images of LIGO glitches and their associated metadata. These glitches are organized by time–frequency morphology into 22 classes for which descriptions and representative images are presented. Results from the application of state-of-the-art supervised classification methods to this dataset are presented in order to provide baselines for future glitch classification work. Standard splitting for training, validation, and testing sets are also presented to facilitate the comparison between different machine learning methods. The baseline methods are selected from both traditional and more recent deep learning approaches. An ensemble framework is developed that demonstrates that combining various classifiers can yield a more accurate model for classification. The ensemble classifier, trained with the standard training set, achieves 98.21% accuracy on the standard test set.

KW - Classification

KW - Dataset

KW - Deep learning

KW - Gravity Spy

KW - Machine learning

KW - aLIGO

UR - http://www.scopus.com/inward/record.url?scp=85042921301&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042921301&partnerID=8YFLogxK

U2 - 10.1016/j.ins.2018.02.068

DO - 10.1016/j.ins.2018.02.068

M3 - Article

VL - 444

SP - 172

EP - 186

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

ER -

Bahaadini S, Noroozi V, Rohani N, Coughlin S, Zevin M, Smith JR et al. Machine learning for Gravity Spy: Glitch classification and dataset. Information Sciences. 2018 May 1;444:172-186. https://doi.org/10.1016/j.ins.2018.02.068